2396 J.C.S. DaltonStructures of Mercury(i1) Halide Adducts with Transition-metal LewisBases. Part l1.l Crystal Structure of the I :3 Complex between Di-carbonyl -.re-cyclopentad ienylcobalt and Mercury( 11) ChlorideBy Ian W. Nowell and D. R. Russell," Department of Chemistry, University of Leicester, Leicester LEI 7RHCrystals of ( ~ T - C ~ H ~ ) ( C O ) ~ C O , ~ H ~ C I ~ are triclinic, a = 10.080, b = 12.572, c = 7.144 8, cc = 87-82, p = 95.67,y = 108.7", and Z = 2 in space group P I . The structure was solved from photographic data by Patterson andFourier methods and refined by least squares t o R 0.095 for 1354 observed reflections. Unlike (x-C5H5) (CO),Co.-HgCI, which i s an adduct, this complex is best formulated as a salt, [(x-C5H5)Co(CO),(HgCI)]+CI-, with twoadditional molecules of mercury(ii) chloride.The cation contains an Hg-Co bond (2.504 * 0.009 A) which issignificantly shorter than that found in the 1 : 1 adduct (2.578 f 0.004 A).EQUIMOLAR amounts of dicarbonyl-x-cyclopentadienyl-cobalt and mercury(I1) chloride react to give the com-plex (x-C,H,) (CO),Co,HgCl,. X-Ray analysis showsthe latter to be a true 1 : 1 adduct containing a metaldonor Co-Hg bond. By the reaction of excess of mer-cury(I1) chloride with either the 1 : 1 adduct or Co-(x-C,H,)(CO),, Cook and Kemmitt obtained ayellow compound, with analysis corresponding to(x-C5H5) (CO),Co,2-7HgCI2. In view of this ratherunsatisfactory formulation, a crystal structure de-termination was undertaken to establish whether thecomplex is a 1 : 1 adduct containing additional mer-cury(~~) chloride molecules in the lattice, or a salt as in theruthenium and osmium complexes: [MII(CO),(PPh,),-(HgCl)] '[HgCIJ-.In the event, the X-ray results show that the crystalsselected for study correspond to [ (55-C5H5) (CO),Co-(HgCl)]+Cl- with two additional molecules of mer-cury(I1) chloride in the crystal lattice.The vibrationparameters of the atoms provide no evidence that theadditional HgCl, sites are occupied in a non-stoicheio-metric manner, and it is possible that the discordancewith the chemical analysis can be accounted for bythe presence of some 1 : 1 adduct in the bulk sample.EXPERIMENTALCrystals were prepared by the reaction of an excess of mer-c u r y ( ~ ~ ) chloride with (x-C5H,) (CO),Co,HgCl,.The yellowprecipitate was recrystallised from acetone [containingmercury(I1) chloride to prevent dissociation of the com-pound] by the addition of petroleum ether (b.p. 40-60 "C).Analysis of the product was still unsatisfactory, but theratio of Co: C: H and Hg: C1 was in agreement withthe formulations Co(x-C5H5) (CO) and HgC1, respectively.Thus the complex was regarded as having the stoicheio-metry (x-C,H,) (CO),Co,nHgCl, and the unsatisfactoryanalysis was attributed to variations in the value of n.Crystals selected for X-ray analysis proved to have thevalue n = 3. Accurate unit-cell dimensions were measuredfrom precession photographs, by use of Zr filtered Mo-K,radiation.Crystal Data.-C,H5Cl,CoHg,0,, M = 994.5, Triclinic,7 No suitable flotation medium could be found to measurethe density of the crystals used for the structure determination.$ Scattering factors for all atoms were taken from ref.5.p For details see Notice to Authors No. 7 in J. Chem. Soc. (A),1970, Issue No. 20 (items less than 10 pp. are sent as full sizecopies).u = 10.08 & 0.02, b = 12-57 f 0.03, G = 7.144 & 0.010 A,a = 87.8 + 0.1, U = p = 95.6 -j= 0.1, y = 108.7 f O-l',887.3 A3, z = 2, D, = 3-72,? ~ ( 0 0 0 ) = 864. m-K,radiation, A = 0-7107 A; ~ ( M o - K , ) = 284 cm-l. Spacegroup Pi confirmed by successful refinement.Data Collection.-Ten reciprocal levels were collectedphotographically by use of Zr filtered Mo-K, radiation.Six layers (hko-5) were collected by the equi-inclinationtechnique using an integrating Nonius camera from acrystal of dimensions 0.05 x 13 x 0.43 mm.A furtherfour reciprocal levels ( O M , lkl, 1201, h 11) were collected froma second crystal of dimensions 0.08 x 0.18 x 0.25 mni byuse of a Supper precession camera. All intensities weremeasured with a Joyce-Loebl integrating microdensito-meter and corrected for Lorentz and polarisation factors.An absorption correction was applied to all data giving,after suitable inter-film and inter-layer scaling, a totalof 1354 independent reflections.Structure A nalysis.-Positions of the three mercuryatoms were found from a three-dimensional sharpenedPatterson map. Structure factors $ for these three atomswere calculated and three cycles of block-diagonal least-squares refinement gave I? 0-28.A difference-Fouriermap enabled the position of six chlorine atoms and cobaltto be found. The atomic scattering factors of mercury,cobalt, and chlorine were corrected for anomalous dis-persion6 and refinement of all ten atoms reduced R to0- 17. -4 further difference-Fourier synthesis revealed theremaining carbon and oxygen atoms, along with electron-density distributed anisotropically about the heavy atoms.Consequently the mercury, cobalt, and chlorine atomswere given anisotropic temperature factors and refinementof all atoms for several cycles, initially with oxygenshaving isotropic temperature factors but finallv withanisotropic values, gave R 0.098. Further refinementwas carried out with the weighting scheme w = l/(a +blKFol + clKFol2) (where a, b, c were 0.935, 0.125, and0.00408) to give a final R value of 0.095.Structure factorsare listed in Supplementary Publication No. SUP 20490(6 pp., 1 microfiche).§ The final positional parameters,with their standard deviations, are listed in Table 1.The final temperature factors are given in Table 2, andrelevant bond distances and angles are listed in Table 3.Part I, I. W. Nowell and D. R. Russell, preceding paper.D. J. Cook and R. D. W. Kemmitt, Chenz. and I n d . , 1966,J. P. Collman and W. R. Roper, Chen?. Conam., 1966, 244.N. W. Alcock's ABSCOR programme, based on the methodof J. De hleulenaer and H. Tompa, Acta Cryst., 1965, 19, 1014.' International Tables for X-Ray Crystallography,' vol.3,Kynoch Press, Birmingham, 1962, p. 201.Ref. 5, p. 203.9461972 2397TABLE 1 TABLE 3Atomic positional parameters (fractional) with theirestimated standard deviations in parenthesesInteratomic distances (A) and angles (") with estimatedstandard deviations of the last figure in parenthesesX Y 2 (4 Distances- 0.01662(3)-0.13999(3)- 0*39982(9)-0*21316(21) -- 0*13743( 19)0-08 142 (3)0.1 7763 (20)0.32635( 17)0*07157( 16)- 0.0025 1 ( 16)- 0.4400(6)-0*4095(6)-0*4298(8)- 0*4039( 6)- 0*4537(8)-0*5855(5)- 0.3425( 8)- 0.5490( 7)- 0.3736 (8)0.05557(3)0-35370(3)0*30647(3)0.26 199 (8)0.1 1338( 18)0.01 164( 19)0.21 793( 20)0-42405( 18)0*26437( 16)0*49726( 15)0.1 809 (2)0*0567(2)0*1968(6)0.1386(5)0.3996( 7)0*3054( 4)0.4383(7)0-2927( 6)0*3678(6)\ I 0*24624( 5 )-0*01950(5)0.47038 (5)0.48141(12)0*07566(30)0*41228(27)0*00564( 36)0*51901(29)0.25429(23)0.1 174( 8)0*6964(8)0-2583( 10)0-6057(9)0*3698( 10)0-4650( 7)0*5678( 10)0-6257( 10)0*7073( 10)- 0.00055(28)2.309 (20)2.314(18)3.086(25)3.141 (20)3.1681 16)3-352( 18)2*336(18)2.332 ( 1 6)3*061(19)2.722 (1 9)3.481 (20)3*506( 17)2*504(9)2.348(16)2 * 839 ( 1 7)3.093( 17)3 * 5 48 (2 7)1.76 (7)1.76(6)2*09( 8)2-1 l(5)2-23(8)2.05 (7)2*08(8)1*04(9)1*17( 8)1*64(9)1*74( 10)1.17(8)1*32(11)1.75( 10)C (3)-C1( 4V3 3.41C (4)-C1(5VI) 3.46(Sum of CH and Clvan der Waals radiiTABLE 2Final temperature factors, with estimated standarddeviations in parentheses(a) Anisotropic temperature factors (A2) *-\tom Bll B 2 2 B33Hg(1)Hg(2)Hg(3)CW)C V )Cl(3)CU4)Cl(5)Cl(6)W)O(2)Xtom B23 Bl, Bl,Hg(2)Hg(3)C1(1)CW)Cl(3)Cl(4)Cl(5)Cl(6)O(1)O(2)Atom B Atom B2*0(9) C(5)1*3(8) C(6)(31)3.3 (1.1) C(7)C(2)(73)C(4)3*02( 14) 3*47( 16) 3*78( 15)160( 12) 3-94( 18) 4.49 ( 1 6)1-03( 11) 4*84( 19) 4.62 ( 16)1 * 33 (33) 2-55(45) 244( 39)3.7(9) 4*0( 1.1) 4*5( 1.0)2*4(8) 5-O( 1.2) 6*6( 1.3)1*8(7) 2.9(9) 6.6 ( 1 * 2)2.4(6) 3*2(9) 2.1(7)1-6( 1.6) 10.1 (3.0) 3*6(2.7)5.2(3.0) 2.2(2-9) 6*9(3*1)c o3.9(9) 4.8 (1.2) 3*4(9)1-1(6) 3.9 (1.0) 4*3(9)0*09( 11) 0-98( 11) 0*36( 12)- 0.24( 13) 0*24( 9) 0.87( 11)0*32(33) 0.50(29) 0*66( 34)0.81 (12) - 0.22(10) - 0.27( 12)c o- 0.6 ( 8) 0.9(8) - 0*4(9)-0.2(8) 1.2(7) 0.3(9)0*5( 1.0) 0*4(8) 0*0(9)-0*3(7) 0.4(6) - 0.3( 7)- 1*8(8) -0*3(7) 0*6(7)0-7(6) 0.4(6) 0*8(7)- OeO(2.7) 0-4( 1.7) 1 4( 2 * 2)- O'O(2.4) 0*4(2*7) 1.4(2*7)(b) Isotropic temperature factors (A2)4*4( 1-2)2*5(9)4.3 ( 1 *2)0*7(7)* In the form exp{-$(B1,h2a*2 + B2,k2b*z + B,,Pc*~ +t In the form exp{-BBsin2 8/A2}.2B,,klb*c* + 2B,,hla*c* + 2B12hka*b*)}.DISC ITS S IONThe structure consists basically of [(x-C,H,)Co-(CO),(HgCl)] + cations, C1- anions, and two distinctmercury( 11) chloride molecules, linked together byHg-Cl interactions to give what is essentially a layerstructure.Figure 1 shows the arrangement withineach layer, neighbouring layers are held together byweak CH - * C1 interactions (Figure 2 and Table 3).3-65)(b) AnglesCl(1)-Hg( l)-Cl(lI) 89.3(2) Co-Hg(3)-C1(3) 99*3(1)Cl( 1)-Hg ( 1 )-Cl( 2) 177.2 ( 1 -5) Co-Hg( 3)-C1(5) 153.5(5)Cl(l)-Hg(l)-C1(21*) 83*4(2) CwHg(3)-C1(6) 116*7(2)C1( 1)-Hg( l)-Cl( 3) 88.1 (2) Co-Hg( 3)-C1( 6v) 106*2( 2)Cl(l)-Hg (l)-Cl( 5 ) 97-3(2) C1( 3)-Hg( 3)-C1(6) 83.1 (2)C1( lI)-Hg( 1)-Cl(2) 88-7( 2) C1(3)-Hg (3)-C1(6) 7 1.3 (2)C1( 1I)-Hg( l)-C1( 211) 83-4( 2) C1(3)-Hg( 3)-C1(gV) 142*0( 4)C1 (l)-Hg ( l)-C1( 3) 88.3 (2) C1(5)-Hg( 3)-C1(6) 89-2 (2)C1( 2)-Hg( 1)-C1( 211) 96.1 (2) C1( 5)-Hg( 3)-C1( 6V) 106*2(3)C1(2)-Hg( l)-Cl( 3) 93.4(2) C1( 6)-Hg(3)-Cl( 6v) 72*7(3)Cl(2)-Hg( 1)-C1(5) 85-6(2)C1(2II)-Hg(l)-C1(5) 73.3(3) Hg( l)-Cl(l)-Hg(2) 85.1(2)C1( 3)-Hg( 1)-C1(5) 79-7( 2) Hg( l)-C1(3)-Hg( 2) 94.4( 2)C1( 1 )-Hg (2)-C1( 3) 78.8( 2) Hg( 2)-C1( 6)-Hg( 2 ) 102.6 (3)C1( 1)-Hg( 2)-C1( 4) 77-7( 2) Hg( 2)-C1( 6)-Hg( 3) 92.4( 2)C1( 1)-Hg( 2)-C1( 5) 80-9( 2) Hg( 3)-C1( 6)-Hg(3v) 107*3(2)Cl( 1)-Hg( 2)-C1( 6)C1( 3)-Hg(2)-C1(4) 156.2 (7) COX( 1)-0 (1) 164(3)C1(3)-Hg( 2)-C1( 5) 80.9 (2) C o x ( 2)-O (2) 177(4)C1( 3)-Hg (2)-C1(6) 8 7-6 (3) C( 3)-C( 4)-C ( 6) 1 lO(4)C1( 4)-Hg( 2)-C1( 5) 90*3( 2) C( 4)-C( 6)-C( 7) 118(4)C1( 4)-Hg( 2)-C1( 6) 104.7 (2) C( 5)-C( 7kC( 6) lOO(3)C1( 4)-Hg( 2)-C1( 6x11) 102-7( 3)C1(5)-Hg(2)-C1(6iII 85-7(2)C1(6)-Hg(2)-C1(6 ) 77-4(3)H m W ) - H g ( 3 ) I I I 8 6 W )1 2 8 q 3)C1( 3)-Hg( 2)-C1(6xxx) 99*7( 3) C( 4)-C( 3)-C( 5 ) 98(3)Roman numerals as superscripts refer t o the followingpositions relative t o the reference molecule at x , y, 2:I -x, -y, -2 IV x , y , -1 + 2 v - x , 1 - y , 1 - 2-1 + x , y , 2I1I11- x , -y, 1 - 2-x, 1 - y , --z VIThe chemistry of the (x-C5H5) (CO),Co,HgCI, systemcan be understood if the following equilibria are assumedto exist in solution :(x-C,H,)Co(CO), + HgC1,(x-CSHJCo (CO),,HgCl, * [ (Z-C~H,) Co(CO),(HgCl)] +C1-(1) (adduct) (2) (salt)Thus the reaction of equimolar amounts of Co(x-C,H,)-(CO), and mercury(I1) chloride gives only a low yieldof the 1 : 1 adduct (1).To prevent dissociation of thelatter upon recrystallisation, an acetone solution con-taining mercury(I1) chloride must be used. The re-action of Co(x-C,H,) (CO), or (x-C,H,) (CO),Co,HgCl,with excess of mercury(I1) chloride forces the equilibri2398 J.C.S.Daltonmore to the right-hand side, thereby favouring form-ation of the salt (2) rather than the adduct. In thesolid state the ionic species (2) is stabilised by an ex-tensive network of Hg-C1 interaction. Although noFIGURE 1 Projection of the unit cell along the a* axis direction.For Hg-Cl distances ~ 2 . 6 A are drawn with dashed lines.clarity, not all Hg-Cl interactions are showndiscrete mercury-containing anion can be identifiedin the lattice, it is reasonable to suppose that suchanions will exist in solutions containing an excess ofmercury(I1) chloride.As the [(x-C~H~)(CO)~CO(H~C~)]+ unit now carries aformal positive charge, one would expect a contractionof the mercury orbitals and a corresponding decreasein the Hg-Co and Hg-Cl bond lengths, compared withthe values found in the 1 : 1 adduct. This is in agree-ment with the experimental results.The Hg-Co bond length is similar to those found inHg[Co(CO)J, (2.498 & 0.007, 2600 -+ 0.007 A) andHg[Co(CO),(PEtJ], (2.499 & 0405 A).'The presence of a formal positive charge on themercury should facilitate a greater donation of electronFIGURE 2 Projection of the unit cell along the b* axisdirectiondensity from the cobalt than in the 1 : 1 adduct.Anyincrease in electron-density donation will be accompaniedby a decrease in back-donation from cobalt to theG. M. Sheldrick and R. N. F. Simpson, J . Chem.SOC. ( A ) ,8 R. F. Bryan and A. R. Manning, Chem. Comm., 1968, 1316.1968, 1006.carbonyl groups, and the higher v(C0) frequenciesfound by Cook and Kemmitt 299 are in agreement withthis. The v(C0) bands at 2069 and 2032 cm-l for thecomplex formulated as (x-C,H,)Co(C0),,2.7HgC12 maybe due to the presence of a certain amount of the1 : 1 adduct. This would also account for the pooranalysis of the complex.The cyclopentadienyl ring is effectively planar.The equation of the best least-squares plane associatedwith the five carbon atoms together with distancesare given in Table 4. The perpendicular distance ofTABLE 4Equation of the best least-squares plane associated withthe five cyclopentadienyl carbon atoms, referred tothe axis system a, b', c*-0.66306X + 0.69560Y + 0.299452 - 8.47799 = 0.0Distances (A) of atoms from the plane: C(3) -0.04, C(4) -0.02,the cobalt atom from this plane (1.66 A) is to be com-pared with the shorter value of 1.70 in the 1 : 1adduct; this observation is also in keeping with therelative effective nuclear charges on the cobalt in thetwo complexes.C(5) 0.08, C(6) 0.06, C(7) 0.09, CO 1-66FIGURE 3 A view of the cation showing additional Hg(3)-ClinteractionsThe co-ordination of Hg(3) in the structure is highlyirregular (Figure 3) with an approximately trigonalbipyramidal geometry.The axial positions areoccupied by the C1- anion [Cl(6')] and a chlorine atom[C1(3)], distance 3.11 and 3-54 A respectively from themercury, with a C1(6')-Hg(3)-C1(3) bond angle of142".The sum of the van der Waals radii for mer-cury lo and chlorine l1 is 3.3 A (RQ = 1.5, Rcl= 1-8 A).Grdenic has suggested lo that some interaction occursup to a value of 1.73 A for R H ~ , giving an upper limitof 3.53 A for REg + Rcl. Even considering this value,the interaction between Hg(3) and Cl(3) must be veryweak at 3.54 A. The four closer chlorine atoms arein a very distorted tetrahedral arrangement aroundHg(3), [Co-Hg(3)-C1(5) and C1(6)-Hg(3)-C1(6') bond0 D. T. Cook. T. L. Dawes, and R. D. W. Kemmitt, 1. Chem.SOG. ( A ) : 1967, lSh7.10 D. Grdenic, Quart. Rev., 1965, 19, 303.11 L. Paulina. ' The Nature of the Chemical Bond,' 3rd edn.,Cornell UniveGity Press, Ithaca, New York1972 2399angles are 153.5 and 72.7'1.This intermediate co-ordination around mercury is very similar to thatfound in the thiophene,l2 [C,H,SHgCI] +C1-, and diethylsulphide,13 [ (Et,S)HgCl] +Cl-,HgCl,, complexes of mer-cury(II)chloride, where salt formation also occurs.The C1- ' anion ' is surrounded by a distorted tetra-hedral arrangement of mercury atoms, such that eachanion bridges two symmetry-related mercury(I1) chlor-ide molecules [C1(3)-Hg(2)-C1(4)]. Some covalent in-teraction exists between the C1- ' anion ' and the mer-cury(11) chloride molecules, evident from distortionsinduced in their molecular geometry. Hg(2) can beconsidered as having a distorted tetrahedral co-ordina-tion of two chlorine atoms [C1(3) and C1(4)] and twochloride anions. Two further chlorines, Cl(5) andC1(1), are within the upper limit of R H ~ + Rcl (3.53 A)for Hg(2) and contribute to asymmetric bridgingbetween Hg(2) and Hg( 1), and Hg(2) and Hg(3).Hg(1) is linked both to Hg(2) and other Hg(1) atomsby asymmetric chlorine bridges, and attains a distortedoctahedral co-ordination of six chlorine atoms.Theoctahedral distortion is similar to that found in mer-cury(I1) chloride itself,l* and the Cl(1)-Hg(l)-Cl(2)molecules are linked together in zigzag chains alongthe (001) direction (Figure 2). These mercury(I1)chloride molecules, in contrast to the C1(3)-Hg(2)-C1(4)l2 C. I. Branden, Arkiv Kemi, 1964, 22, 496.13 C. I. Branden, Arkiv Kemi, 1964, 22, 83.l4 D. Grdenic, Arkiv Kemi, 1960, 22, 14.molecules, are only slightly distorted. The two shortHgCl distances are 2-31 A, which although longerthan that found in free mercury(I1) chloride (2.26 A)are similar to the values found13 (2.30 and 2.33 A)in the unsubstituted HgCl, of the diethyl sulphideadduct, [(Et2)SHgCl+]C1-,HgCl,.The formation of salts rather than discrete adductsmay well occur in the reaction of other transition-metal complexes with mercury(I1) chloride. ThusCollman and Roper3 have formulated the rutheniumand osmium adducts, M(CO),(PPh,),,ZHgX, (X =CI, Br, or I) as salts: [MII(CO),(PPh,),(HgX)]+[HgXJ-.Complexes such as Fe(CO),(P(OPh),),,SHgC1, l5 and(x-C6H3Me,)M(C0),,2HgC1, (M = Cr or Mo) l6 may wellbe salts also. However, if similar equilibria are presentas that postulated in the present system, detailed struc-tural analysis would be necessary to confirm theseformulations.We thank the S.R.C. for a grant (to I. W. N.), for equip-ment and for time on the Atlas computer at Chilton,Berkshire. Calculations were also carried out on theUniversity of Leicester Elliott 4130 and 803B computersusing programs written by us.[1/1448 Received, 13th August, 19711l5 D. M. Adams, D. J. Cook, and R. D. W. Kemmitt, Chenz.l6 K. Edgar, B. F. G. Johnson, J. Lewis, and S. B. Wild,Comm., 1966, 103.J . Chem. SOG. ( A ) , 1968, 2861